A recent study conducted by MIT reveals that the availability of the assistive chatbot ChatGPT has resulted in a 40 percent reduction in the time required by workers to complete certain writing tasks, along with an 18 percent improvement in output quality. The findings were published in a press release issued by the institution on Friday.
Shakked Noy, a PhD student in MIT’s Department of Economics and co-author of the paper, emphasized the significant impact of generative AI on white-collar work. However, Noy also acknowledged that it is still too early to fully comprehend the overall implications and effects of this technology on society.
Noy believes that this emerging technology can enhance worker productivity and have a positive net effect on the economy. From an economic perspective, productivity is considered a crucial aspect of technological advancements, allowing for more efficient economic output.
The study involved 453 professionals, including marketers, grant writers, consultants, data analysts, human resource professionals, and managers, who were assigned two writing tasks relevant to their respective occupations. Experienced professionals in the same fields evaluated each submission without knowing which ones were assisted by ChatGPT.
The results indicated that participants who used ChatGPT completed their tasks approximately 11 minutes faster compared to the control group. Additionally, the average quality evaluations of their work increased by 18 percent.
Notably, the study also revealed that workers exposed to ChatGPT during the experiment were twice as likely to continue using it in their actual jobs two weeks later.
While acknowledging the time-saving benefits of ChatGPT, Noy acknowledged that the real-world implementation requires additional time for fact-checking and writing prompts.
Moving forward, the researchers aim to extend their understanding of generative AI’s impact on the economy based on the insights gained from ChatGPT’s influence.
The researchers acknowledge that there are various other factors that will affect wages, employment, and shifts across sectors. These factors necessitate additional evidence beyond the scope of their paper. Nonetheless, given the substantial time saved and quality improvements observed in their study, the researchers consider generative AI to be a potentially revolutionary development, particularly for certain types of work.
Furthermore, the researchers aim to assess how society should respond to the widespread adoption of generative AI. They emphasize the need to analyze future research findings to determine the appropriate policy adjustments required to address the economic and political implications. For instance, understanding whether generative AI will lead to increased wages for lower-paid workers or exacerbate wage inequality by benefiting high earners is crucial for developing appropriate responses. The economic and political effects downstream should be thoroughly examined to make informed decisions.
